I fumbled for a few years before worming my way into a middle-ranked graduate department of Neuroscience. Learned how to work an MRI and how to hack scripts in Matlab. Realized I could think in maths/stats even though none of my transcripts suggested I could (my only undergrad math course was calculus, I took it twice, and the second time I got a 67). I went to a conference in 2012 and saw a Harvard Business Review article on big data in the airport on my way home. I thought to myself "no one is going to care about my MRI research but I bet I can become a data scientist without a PhD".

So I found a lab-tech job post-MSc for a bit and then I became a "research methods specialist" and worked in a research hospital for 4 years. I learned a lot about high-performance computing, the kinds of models used for pre-clinical trials, and methods development for clinical research. I ended up on a lot of papers. My CV now rivals many PhD grads I know.

Then through some networking and persistence, I find myself at the tail end of a professional MSc in computer science at a top-ranked machine learning institute. I have successfully survived both core CS coursework and graduate courses in reinforcement learning and deep learning. I'm currently working on multiple deep learning research projects and getting to know some really interesting people doing biomedical startups.

10 years ago I was living in my dad's basement, a psych grad who couldn't math and showed pretty much zero aptitude for these kinds of endeavors. I never would have believed that I would be in the kind of program I am in now. I'm only 32, and my story isn't done, but I'm sure if I make it onto any list, it will be one for people over the age of 35.

Great to see you being an ace though! I hope you keep moving forward.

If you want to get involved in research projects (which I do think helps you land more creative jobs) I would suggest working your hard-core software dev background when you pitch potential collaborators. You would be very surprised to see the quality of code coming out of some researchers because there is little incentive to be a good software engineer in academia. The normal way to participate in research is to be in a PhD program but there are many other ways so long as you can work closely with people who come from a more traditional research trajectory.

Are companies in geospatial or government aerospace groups hiring people with your skills? Might be a first step?

At my school, I was required to take algos and data structures in addition to the program requirements. This was to make up for the fact that I wasn't already a CS grad. Way less time effort and money than a full CS degree! So maybe you can find a similar arrangement?

From my experience, if you're going to pay for more school, you should try to get a masters. A prof once gave me the same advice "it's more work (i.e., a BSc is typically 3-4 years) and in the end you will just have another bachelors". The nice thing about MSc programs is they tend to be more project-based and focus on higher-level concepts so you learn more useful / high-level stuff than in undergrad IMHO.

Good luck! There's always an alternate path :)

Before the switch, I was used to working insane hours, and I had little money (due the failed restaurant). So I was hungry, no family to support, and in a perfect situation to go for broke.

Today? I could never make the same changes. Supporting a family, and a wife with a career of her own. A mortgage, and retirement in a decade. All these things limit the amount of change you can (or are willing to) take. So I put my curiosity into hobbies etc.

Also from Australia, and we are a pretty optimistic bunch in polite company. My opinion is that people are increasingly more likely to settle for the life they have already, so they don't try big things, not that they get less likely to succeed at them.

As someone else pointed out though, we aren't in the silicon valley rat race. We have a much less turbulent and ephemeral economy to work in. I think if you take on real problems you can see long term success here, it's just not going to be a billion dollar venture funded exit.

What exactly are you hoping to hear? Because, among other things, I walked away from a National Merit Scholarship. I think I did the right thing and I think I'm doing important stuff, but the world still sees me as a big fat nobody.

So what exactly are you looking for? "Life advice"? Only to hear from people with some kind of substantial public recognition who got started late?

You'll like his work if you enjoy literary fiction with a formalist and philosophical bent. Fairly calm and airy.

My degrees were in economics and arts, which I swapped to from CS explicitly because i wanted to learn things I didn't know, not get good grades on things I did. I worked security night shift during uni which let me read and study whatever I wanted while being paid for it.

From there I took an relatively unskilled level job doing order processing with a legal publishing firm. What i consider my first 'real' program was written there. We did all our processing through this crappy slow manual interface with manual data entry, menu selection, and second long pauses between menus and pages. I hypothesized you could map out each menu and form in the entire application via unique pixel patterns grabbed from each screen and wrote a program that could get from any place in the software to any other and knew where in the menu system it was. Then I wrote a simple spreadsheet interface with some simple memory (since I observed huge portions of our entry was the same data/orders from the same customers, so why do we re-enter the same details multiple times?). I hooked it all together and could thus spend the first hour or two of the day filling out the data spreadsheet: the program could then read the spreadsheet and essentially do my job for me. It new how to interface with my workstation, knew where it was, where to go for every order type, didn't get bored or make mistakes and entered everything infinitely faster than a human typist. I have a real soft spot for that program, watching it zoom the mouse around the screen and do everything automatically still makes me feel proud.

From there I moved to a performance manager job where I implemented algorithms to try to back-engineer government rating systems and use our sample of the market to offer predictions before government sources were published.

Then I moved into a research/policy position.

I used this to get work at the national stats bureau: where I volunteered to work on the hard programming problems no one else wanted. I then transferred into the analytics specialist section therein where I started my work and research on data linking.

Then I used that to move cities and got employed by a bank as an analytics specialist.

Then I left there and am now with my current employer.

I'm just over 35, live where I want to live, never stopped learning either programming or my other fields, haven't had to work for any job I consider unethical, have almost paid off my home, started a young family, and am gainfully employed with experience in philosophy, economics, stats, analytics, computer science, security, and publishing :p

I had failed grades and often barely handed in assignments for uni, and now I'm relatively confident I could run circles around most academics.

Anyone who says your best work happens in your 20s is an ignorant fool. Knowledge compounds. I was relatively smart during my 20s, but I was nothing compared to what I am now in my 30s.

I expect my 40s will be even better...